An artificial intelligence (AI) device includes a sensing unit configured to collect operation log including information on an external environment factor and an operation state of an AI device, a memory configured to store data corresponding to the operation log, and a processor configured to provide the data corresponding to the operation log to an AI model, to acquire information about whether the AI device corresponds to a normal range or a malfunction symptom range, and to perform control based on the acquired information.
Legal claims defining the scope of protection, as filed with the USPTO.
2. The AI device of claim 1, wherein the AI model is a neural network trained by labeling the information on the normal state or the malfunction symptom state to training data corresponding to the operation log.
3. The AI device of claim 1, wherein the processor provides the data corresponding to the operation log to the AI model and acquires classification result information about a state corresponding to the operation log among at least one normal state or at least one malfunction symptom state based on a classification value that is output using the provided data by the AI model.
4. The AI device of claim 1, wherein the processor deletes the data corresponding to the operation log stored in the memory upon determining the AI device corresponds to the normal state.
5. The AI device of claim 1, wherein the processor stores the data corresponding to the operation log stored in the memory as an unusual log in the memory upon determining the AI device corresponds to the malfunction symptom state.
6. The AI device of claim 5, wherein the processor allows a request for access to the unusual log stored in the memory when there is malfunction receipt corresponding to the acquired information.
7. The AI device of claim 1, wherein the processor labels the information on the normal state to the data corresponding to the operation log and provides the labeled information to the AI model when there is not malfunction receipt corresponding to the acquired information within a preset time.
8. The AI device of claim 1, wherein the processor labels the information on the malfunction symptom state to the data corresponding to the operation log and provides the labeled information to the AI model when there is malfunction receipt corresponding to the acquired information within a preset time.
9. The AI device of claim 1, wherein the at least one sensor includes a dust sensor configured to collect the fine dust concentration as the information on the external environment factor of the AI device, and a motor sensor configured to collect the revolutions per minute of the motor as the information on the operation state of the AI device.
10. The AI device of claim 1, wherein the processor provide data of a time sector corresponding to the operation log to the AI model and acquires the information about whether the AI device corresponds to the normal state or the malfunction symptom state.
11. The AI device of claim 10, wherein the data corresponding to the operation log collected in time-series is separated to a predetermined time sector.
13. The method of claim 12, wherein the AI model is a neural network trained by labeling the information on the normal state or the malfunction symptom state to training data corresponding to the operation log.
14. The method of claim 12, wherein the acquiring the information includes providing the data corresponding to the operation log to the AI model and acquiring classification result information about a state corresponding to the operation log among at least one normal state or at least one malfunction symptom state based on a classification value that is output using the provided data by the AI model.
20. The method of claim 12, wherein the acquiring the information includes providing data of a time sector corresponding to the operation log to the AI model and acquiring the information about whether the AI device corresponds to the normal state or the malfunction symptom state.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
August 30, 2019
September 20, 2022
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.